Multi-Signal ERP Graphs for Predictive & Prescriptive Supply Chain Resilience
DOI:
https://doi.org/10.63282/3050-922X.IJERET-V5I3P118Keywords:
Supply-Chain Resilience, Enterprise Resource Planning (ERP), Graph Neural Networks, Multi-Signal Fusion, Prescriptive Analytics, Early-Warning Systems, ESG ComplianceAbstract
Modern supply-chain networks are becoming more susceptible to disruptions due to geopolitical shifts, changing demand, and pressures related to ESG compliance. Although Enterprise Resource Planning (ERP) systems provide transactional visibility, they fall short in offering substantial predictive capabilities, and independent AI modules typically fail to enable prescriptive decision-making. This study introduces a Multi-Signal ERP Graph (MSEG) framework that merges various enterprise, IoT, and ESG data into a unified graph model to facilitate predictive risk assessment and provide prescriptive strategies for resilience. By utilizing a hybrid Graph Neural Network (GNN) along with a constrained optimization model, MSEG evaluates the probability of disruptions and recommends mitigation strategies while taking into account cost and sustainability constraints. Simulation experiments carried out on synthetic supply chains consisting of 300 nodes demonstrate an average enhancement of 3.2 days in detection lead time, an increase of 8.6% in fill rates, and a 76% decrease in ESG violations compared to conventional ERP alerts. The proposed system improves both predictive and prescriptive analytics, leading to a more efficient decision-making process for resilient and sustainable supply chain management.
References
[1] C. Smyth, "Artificial intelligence and prescriptive analytics for supply chain resilience," Int. J. Prod. Res., vol. 62, no. 4, pp. 1201–1218, Apr. 2024.
[2] T. M. Choi, S. S. Chiu, and C. W. Chan, "Optimization models for supply chain management: A review," Comput. Oper. Res., vol. 180, p. 106331, May 2024.
[3] G. Zheng and A. Brintrup, "An analytics-driven approach to enhancing supply chain visibility with graph neural networks," arXiv preprint arXiv:2403.07231, Mar. 2024.
[4] F. Qi, L. Zhang, K. Zhuo, and X. Ma, "Early warning for manufacturing supply chain resilience based on improved grey prediction model," Sustainability, vol. 14, no. 20, p. 13125, Oct. 2022.
[5] S. Yang, K. Ikeuchi, and Y. Okuma, "Post-hazard supply chain disruption: Predicting firm-level sales using graph neural network," Int. J. Disaster Risk Reduct., vol. 110, p. 104664, Jul. 2024.
[6] S. A. H. Shekarabi et al., "Supply chain resilience: A critical review of risk, metrics and methods," J. Bus. Logist., vol. 45, no. 1, pp. 88–110, Jan. 2024.
[7] Qi, F., Zhang, L., Zhuo, K., & Ma, X. (2022). Early warning for manufacturing supply chain resilience based on improved grey prediction model. Sustainability, 14(20), 13125. https://doi.org/10.3390/su142013125P. Li et al., "Digital transformation and supply chain resilience: The role of cloud-based ERP," Technol. Forecast. Soc. Change, vol. 201, p. 123250, Apr. 2024.
[8] B. Wang and Y. Xue, "Spatio-temporal graph neural networks for industrial chain resilience," arXiv preprint arXiv:2308.16836, Aug. 2023.
[9] Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, "A comprehensive survey on graph neural networks," IEEE Trans. Neural Netw. Learn. Syst., vol. 32, no. 1, pp. 4–24, Jan. 2021.
[10] A. Alaoua et al., "Intelligent early warning system for supplier delays using machine learning," Machines, vol. 8, no. 5, p. 124, May 2024.
[11] K. Xu et al., "Representation learning on large-scale supply chain graphs," in Proc. Int. Conf. Learning Representations (ICLR), May 2024.
